ruby_llm 2.0.0.rc3 → 2.0.0.rc4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: bf1e2bc13df637c93268a6ee20e52c704c47f942be416c4d6728882ae2cc35bf
-  data.tar.gz: b63c7e4eef810843ad1bcb91a8ad55413f90764ba7e984d310c8a6b4bec3bec0
+  metadata.gz: c047801a2b7a5a642296f0ae84e81ab29b641353ffe8618b3ece44db0b0a3849
+  data.tar.gz: 05e4624f14a8e2045408dd21007758612f0644522bae4e250902fd60d66955e1
 SHA512:
-  metadata.gz: 030126750af8dea0303a81093d415e361fdfb7335620d8e84da3fd024acba2e5323bb7fcbd47530a6a3d5cbc24ba3c2f771db431aabd8011502bd19b9b68aa71
-  data.tar.gz: c783cfc4c121db83f4e05867893b7b822ffeddb53010eb22a5053d37eca7a539ee26dc6181b669ac5b2fac856464a18cc8929303127c1e5ae1f1fe28f9c1f373
+  metadata.gz: 6e73e5f6818e70c607c3cabfc6e0dfb19f9271491375b71279e0a205f2ecff8fe0051c691c51230644d32f7ad8f377b8f8bddfcae919e33ff64015ccd9a942e7
+  data.tar.gz: fd4285734764621dc2cab77579b8c713aaff907930c1b6849472c470839b2a75e6975832baf4e43adeedd1d3326b92af816af4a63242bdfa2585fab2b72fdc58

data/.rdoc_options

@@ -17,8 +17,8 @@ exclude:
   - \.json$
 footer_content:
   Documentation:
-    Guides: https://rubyllm.com/next/
-    Upgrading: https://rubyllm.com/next/upgrading/
+    Guides: https://rubyllm.com/
+    Upgrading: https://rubyllm.com/upgrading/
   Resources:
     GitHub: https://github.com/crmne/ruby_llm
     Issues: https://github.com/crmne/ruby_llm/issues

data/README.md

@@ -7,7 +7,9 @@
 
 <strong>Build AI features the Ruby way</strong>
 
-<p>The Ruby-native AI framework. Build with chats, tools, agents, images, audio, and video through one consistent API, in plain Ruby or Rails.</p>
+<p>RubyLLM is the Ruby-native AI framework. Work with models, tools, and agents through one consistent API, in plain Ruby or Rails.</p>
+
+[Website](https://rubyllm.com/) · [Getting Started](https://rubyllm.com/getting-started/) · [What's New in 2.0](https://rubyllm.com/whats-new-in-2-0/)
 
 Battle tested at [<picture><source media="(prefers-color-scheme: dark)" srcset="https://chatwithwork.com/logotype-dark.svg"><img src="https://chatwithwork.com/logotype.svg" alt="Chat with Work" height="30" align="absmiddle"></picture>](https://chatwithwork.com) - *Fully private work AI*
 
@@ -24,26 +26,19 @@ Battle tested at [<picture><source media="(prefers-color-scheme: dark)" srcset="
 
 ---
 
-Work with OpenAI, xAI, Anthropic, Google, AWS, local models, and more. Seventeen providers are built in, and you can connect an OpenAI-compatible endpoint directly.
-
-## Build a working Ruby AI chat in two minutes
-
-https://github.com/user-attachments/assets/65422091-9338-47da-a303-92b918bd1345
-
-## Why RubyLLM?
+## 17 providers. One Ruby API.
 
-Use the same Ruby methods across providers. Add files to a conversation, give an agent tools, generate media, or build a search feature with embeddings and reranking. Read response text, generated files, and usage through Ruby objects.
+Build with the models you want. Move between hosted and local providers without rewriting your application, or connect an OpenAI-compatible endpoint.
 
-In Rails, the API works on your own Chat and Message records, with Active Storage attachments, Hotwire streaming, and background jobs. RubyLLM maintains the supporting model registry, tool calls, usage ledger, and batches. A handful of small dependencies keeps it easy to bring into an existing application.
+[Browse models and pricing](https://rubyllm.com/available-models/) · [Connect a provider](https://rubyllm.com/configuration-providers/)
 
-## Show me the code
+## Start with one line. Add files, tools, and agents
 
-These examples use **2.0.0.rc3**. Follow [Getting Started](https://rubyllm.com/next/getting-started/) to install it and configure the providers you want to try.
+These examples use **2.0.0.rc4 (prerelease)**. Follow [Getting Started](https://rubyllm.com/getting-started/) to install it and configure the providers you want to try. For 1.x, use the [stable-version docs](https://rubyllm.com/v1/).
 
 ```ruby
-# Just ask questions
-chat = RubyLLM.chat
-chat.ask "What's the best way to learn Ruby?"
+# Just ask
+RubyLLM.chat.ask "What's the best way to learn Ruby?"
 ```
 
 ```ruby
@@ -155,7 +150,9 @@ response = chat.with_schema(ProductSchema).ask "Analyze this product", with: "pr
 response.parsed
 ```
 
-## Features
+## A complete AI framework for Ruby
+
+Agents, workflows, RAG, images, audio, and video. Built in, with usage tracking and Rails integration to bring them into your app.
 
 * **Chat:** Conversational AI with `RubyLLM.chat`
 * **Vision:** Analyze images and videos
@@ -195,7 +192,7 @@ response.parsed
 Install the 2.0 release candidate:
 
 ```bash
-bundle add ruby_llm --version 2.0.0.rc3
+bundle add ruby_llm --version 2.0.0.rc4
 ```
 
 Configure a provider in your script, or in `config/initializers/ruby_llm.rb` in Rails:
@@ -208,9 +205,13 @@ RubyLLM.configure do |config|
 end
 ```
 
-Configure the other providers used by the examples as needed: Gemini for files, xAI for video, Mistral for OCR, and Cohere for reranking. [Getting Started](https://rubyllm.com/next/getting-started/) shows each setup beside its example. If your app uses 1.16, follow the [upgrade guide](https://rubyllm.com/next/upgrading/) before deploying 2.0.
+Configure the other providers used by the examples as needed: Gemini for files, xAI for video, Mistral for OCR, and Cohere for reranking. [Getting Started](https://rubyllm.com/getting-started/) shows each setup beside its example. If your app uses 1.16, follow the [upgrade guide](https://rubyllm.com/upgrading/) before deploying 2.0.
 
-## Rails
+## Feels at home in Rails
+
+Save conversations with Active Record and stream replies with Hotwire. The generators give you a working chat UI. Watch the two-minute demo.
+
+https://github.com/user-attachments/assets/65422091-9338-47da-a303-92b918bd1345
 
 ```bash
 # Install Rails Integration
@@ -231,11 +232,21 @@ chat = Chat.create! model: "gpt-5.6-luna"
 chat.ask "What's in this file?", with: "report.pdf"
 ```
 
-Visit `http://localhost:3000/chats` for a ready-to-use chat interface!
+Start your Rails server and visit `http://localhost:3000/chats` to try the chat interface. See the [Rails guide](https://rubyllm.com/rails/) for persistence, streaming, and background jobs.
+
+## AI coding assistants
+
+Give your coding assistant the RubyLLM API and documentation that match your application. From your application directory, install the skill packaged with your gem:
+
+```bash
+npx skills add "$(bundle show ruby_llm)" --skill rubyllm
+```
+
+Choose your coding assistant and installation scope when prompted. See [AI Coding Assistants](https://rubyllm.com/ai-coding-assistants/) for setup and updates.
 
 ## Documentation
 
-[Guides](https://rubyllm.com/next/getting-started/) · [API reference](https://rubyllm.com/next/api/) · [Models](https://rubyllm.com/available-models/) · [Upgrading](https://rubyllm.com/next/upgrading/)
+[Guides](https://rubyllm.com/getting-started/) · [API reference](https://rubyllm.com/api/) · [Models](https://rubyllm.com/available-models/) · [Upgrading](https://rubyllm.com/upgrading/) · [1.x docs](https://rubyllm.com/v1/)
 
 ## Contributing
 

data/lib/generators/ruby_llm/upgrade/online_copy_migration/data.rb

@@ -63,6 +63,15 @@ module RubyLLM
           end
         end
 
+        def discard_incomplete_tool_calls
+          source = records(configuration.fetch('tool_call_table'))
+          result_key = configuration.fetch('tool_call_foreign_key')
+          results = messages.where.not(result_key => nil).select(result_key)
+          owners = legacy_messages.select(messages.primary_key)
+          source.where(configuration.fetch('message_foreign_key') => owners)
+                .where.not(source.primary_key => results).delete_all
+        end
+
         def catch_up(passes: nil)
           sync_models
           pass = 0

data/lib/generators/ruby_llm/upgrade/online_copy_migration.rb

@@ -47,7 +47,7 @@ module RubyLLM
         end
       end
 
-      def finish
+      def finish(discard_incomplete_tool_calls: false)
         with_migration_lock do
           state = states.first!
           return if state.active_version == 2 && state.status == 'active' && data.finished?
@@ -55,6 +55,10 @@ module RubyLLM
           raise 'Run the copy backfill before finish' unless data.completed?
 
           pause(state, from: 1, statuses: %w[preparing active finishing])
+          if discard_incomplete_tool_calls
+            discarded = data.discard_incomplete_tool_calls
+            ::ActiveRecord::Migration.say "Discarded #{discarded} incomplete legacy tool calls"
+          end
           data.catch_up
           data.verify
           verify_completed_work

data/lib/generators/ruby_llm/upgrade/templates/backfill_v2_data.rb.tt

@@ -1,15 +1,17 @@
 # frozen_string_literal: true
 
-require 'generators/ruby_llm/upgrade/legacy_content_sql'
 <% if copy_mode? -%>
-
 require 'generators/ruby_llm/upgrade/upgrade_migration'
+<% else -%>
+require 'generators/ruby_llm/upgrade/legacy_content_sql'
 <% end -%>
 
 class <%= backfill_migration_class_name %> < ActiveRecord::Migration<%= migration_version %>
   disable_ddl_transaction!
 
+<% unless copy_mode? -%>
   BATCH_SIZE = 10_000
+<% end -%>
   PROGRESS_TABLE = :ruby_llm_v2_backfills
 
   def up
@@ -46,6 +48,7 @@ class <%= backfill_migration_class_name %> < ActiveRecord::Migration<%= migratio
     raise 'The RubyLLM 2.0 upgrade is already finished. Do not run the backfill again.'
   end
 
+<% unless copy_mode? -%>
   def backfill_required_defaults
     {
       <%= chat_table_name %>: {cancelled: false},
@@ -453,6 +456,7 @@ class <%= backfill_migration_class_name %> < ActiveRecord::Migration<%= migratio
     yield
   end
 
+<% end -%>
   def progress_records = migration_record(PROGRESS_TABLE)
 
   def migration_record(table)

data/lib/generators/ruby_llm/upgrade/templates/finish_v2_upgrade.rb.tt

@@ -13,7 +13,7 @@ class <%= finish_migration_class_name %> < ActiveRecord::Migration<%= migration_
   def up
 <% if copy_mode? -%>
     RubyLLM::Generators::UpgradeMigration.for.verify_settings(<%= copy_upgrade_settings.inspect %>)
-    RubyLLM::Generators::UpgradeMigration.for.finish do
+    RubyLLM::Generators::UpgradeMigration.for.finish<% if discard_incomplete_tool_calls? %>(discard_incomplete_tool_calls: true)<% end %> do
       verify_completed_backfills
       enforce_required_defaults
       relax_legacy_message_constraints
@@ -113,6 +113,7 @@ class <%= finish_migration_class_name %> < ActiveRecord::Migration<%= migration_
     raise "RubyLLM 2.0 backfills are incomplete: #{missing.join(', ')}"
   end
 
+<% unless copy_mode? -%>
   def verify_message_content
     return unless column_exists?(:<%= message_table_name %>, :content_raw)
 
@@ -220,17 +221,22 @@ class <%= finish_migration_class_name %> < ActiveRecord::Migration<%= migration_
 <% end -%>
   end
 
+<% end -%>
   def with_upgrade_safety(&)
     return safety_assured(&) if respond_to?(:safety_assured, true)
 
     yield
   end
 
-  def message_value(column) = "legacy_messages.#{quote_column(column)}"
   def column_definition(table, column) = connection.columns(table).find { |candidate| candidate.name == column.to_s }
+<% unless copy_mode? -%>
+  def message_value(column) = "legacy_messages.#{quote_column(column)}"
   def quoted_primary_key(table) = quote_column(connection.primary_key(table))
   def quote_table(table) = connection.quote_table_name(table)
+<% end -%>
+<% if postgresql? || !copy_mode? -%>
   def quote_column(column) = connection.quote_column_name(column)
+<% end -%>
 
   def migration_record(table)
     Class.new(ActiveRecord::Base) do

data/lib/generators/ruby_llm/upgrade/templates/ruby_llm_upgrade.rb.tt

@@ -15,7 +15,7 @@ module RubyLLMUpgrade # :nodoc: all
       self.enumerate_columns_in_select_statements = true
       default_scope { RubyLLMUpgrade.visible(self) }
       after_initialize { @ruby_llm_upgrade_epoch = RubyLLMUpgrade.epoch }
-      around_save :guard_ruby_llm_upgrade, prepend: true
+      around_save :guard_ruby_llm_upgrade
       around_destroy :guard_ruby_llm_upgrade, prepend: true
       before_save { RubyLLMUpgrade.prepare_model(self) if ruby_llm_upgrade_kind == :chat }
     end

data/lib/generators/ruby_llm/upgrade/templates/upgrade_initializer.rb.tt

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
-require 'active_record/base'
 require Rails.root.join('app/models/concerns/ruby_llm_upgrade')
 
 Rails.application.config.to_prepare do
+  require 'active_record/base'
   tool_call = if Gem::Version.new(RubyLLM::VERSION).segments.first >= 2
                 RubyLLM::ActiveRecord::ToolCall
               else

data/lib/generators/ruby_llm/upgrade/upgrade_generator.rb

@@ -26,6 +26,8 @@ module RubyLLM
                            desc: 'Generate one upgrade phase; cleanup runs in a later deployment'
       class_option :mode, type: :string, enum: %w[rename copy], default: 'rename',
                           desc: 'Copy prepares and backfills online, with protected 1.16 rollback until cleanup'
+      class_option :discard_incomplete_tool_calls, type: :boolean, default: false,
+                                                   desc: 'Discard all incomplete legacy tool calls at copy finish'
 
       argument :model_mappings,
                type: :array,
@@ -39,6 +41,7 @@ module RubyLLM
       end
 
       def create_migration_files
+        validate_cleanup_options
         parse_model_mappings(allowed_types: MODEL_MAPPING_TYPES, defaults: MODEL_MAPPING_DEFAULTS)
         say_status :models, resolved_model_mappings
 
@@ -132,6 +135,14 @@ module RubyLLM
       private
 
       def copy_mode? = options[:mode] == 'copy'
+      def discard_incomplete_tool_calls? = options[:discard_incomplete_tool_calls]
+
+      def validate_cleanup_options
+        return unless discard_incomplete_tool_calls?
+        return if copy_mode? && options[:phase].in?([nil, 'finish'])
+
+        raise Thor::Error, 'Discard incomplete tool calls only with --mode copy and the finish phase'
+      end
 
       def copy_upgrade_settings
         {

data/lib/generators/ruby_llm/upgrade/upgrade_migration.rb

@@ -248,7 +248,8 @@ module RubyLLM
         calls = records(:ruby_llm_tool_calls).where(message_type: configuration.fetch('message_class'), result_id: nil)
         return unless calls.exists?
 
-        raise 'Finish pending tool calls and approvals before switching versions'
+        raise 'Finish pending tool calls and approvals before switching versions. ' \
+              'For abandoned legacy calls, see https://rubyllm.com/upgrading/#incomplete-tool-calls'
       end
 
       def reconcile

data/lib/ruby_llm/batch.rb

@@ -101,7 +101,7 @@ module RubyLLM
         chats = normalize_chats(records)
 
         provider = shared_provider(chats)
-        payload = { provider: provider.slug, provider_class: provider.class.display_name, requests: chats.size }
+        payload = { provider: provider.slug, provider_class: provider.name, requests: chats.size }
         RubyLLM.instrument('batch.ruby_llm', payload, config: provider.config) do |event|
           requests = chats.each_with_index.map do |chat, index|
             { custom_id: index.to_s, model: chat.model.id, payload: chat.render }
@@ -120,7 +120,7 @@ module RubyLLM
         end
 
         provider = shared_provider(requests)
-        payload = { provider: provider.slug, provider_class: provider.class.display_name, requests: requests.size }
+        payload = { provider: provider.slug, provider_class: provider.name, requests: requests.size }
         RubyLLM.instrument('batch.ruby_llm', payload, config: provider.config) do |event|
           lines = requests.each_with_index.map do |request, index|
             { custom_id: index.to_s, model: request.model.id, payload: request.render, text: request.text }

data/lib/ruby_llm/chat.rb

@@ -1003,7 +1003,7 @@ module RubyLLM
       {
         chat: self,
         provider: @provider.slug,
-        provider_class: @provider.class.display_name,
+        provider_class: @provider.name,
         model: @model.id,
         model_info: @model,
         input_messages: messages.dup,
@@ -1347,7 +1347,7 @@ module RubyLLM
       payload = {
         chat: self,
         provider: @provider.slug,
-        provider_class: @provider.class.display_name,
+        provider_class: @provider.name,
         model: @model.id,
         model_info: @model,
         tool: tool,

data/lib/ruby_llm/embedding.rb

@@ -117,7 +117,7 @@ module RubyLLM
 
       payload = {
         provider: provider_instance.slug,
-        provider_class: provider_instance.class.display_name,
+        provider_class: provider_instance.name,
         model: model_id,
         model_info: model,
         input: text,

data/lib/ruby_llm/image.rb

@@ -75,7 +75,7 @@ module RubyLLM
       empty_tokens = Tokens.new
       payload = {
         provider: provider_instance.slug,
-        provider_class: provider_instance.class.display_name,
+        provider_class: provider_instance.name,
         model: model.id,
         model_info: model,
         prompt: prompt,